旧版入口
学术交流
2017年6月19日Defeng Sun 教授学术报告
上传时间:2017-06-14 作者:杭电理学院 浏览次数:
报告题目: First and Second Order methods for Convex and Nonconvex Statistical Regression Models  
报告人:Defeng Sun   教授
报告内容:
Statistical regression models of big scales are ubiquitous in machine learning, statistics, finance, signal processing, imaging science, geophysics and many other areas. Concerned with the huge computational burdens of the interior point methods for solving these big scale problems, convex or nonconvex, many researchers and practitioners tend to believe that the first order methods such as the accelerated proximal gradient methods and the alternating direction methods of multipliers are the only options for the rescue. While these first order methods have enjoyed successful stories in some interesting applications, they also encounter enormous numerical difficulties in dealing with many real data problems of big scales even only with a low or moderate solution quality. New ideas for solving these problems are highly sought both in practice and academic research. In this talk, we shall first demonstrate how the second order sparsity property exhibited in big sparse regression models can be intelligently explored to overcome the mentioned difficulties either in IPMs or in the first order methods. One critical discovery is that the second order sparsity allows one to solve sub-problems at costs even lower than several first order methods. For the purpose of illustration, we shall present a highly efficient and robust semismooth Newton based augmented Lagrangian method for solving various lasso and support vector machine models. Finally, we shall discuss about a roadmap on the choices of algorithms for the regression models of different features.
报告人简介:
    Defeng Sun is Professor at Department of Mathematics and Risk Management Institute, National University of Singapore. His main research interest lies in large scale matrix optimization and statistical learning. Currently he serves as associate editor to Mathematical Programming, both Series A and Series B, SIAM Journal on Optimization and others.
报告时间:2017年6月19日 15:00 -16:00
报告地点:理学院学术报告厅(六教南528)
                               欢迎感兴趣的老师和同学们参加!
Copyright © 2014 杭州电子科技大学 版权所有
技术支持 : 捷点科技